Noninvasive IDH1 genotype prediction in grade Ⅱ/Ⅲ  gliomas based on conventional MR images: a transfer learning strategy

Purpose: To evaluate the performance of transfer learning with CNNs in predicting IDH1 genotype. Method and Materials: AlexNet, GoogLeNet, ResNet and VGGNet were pre-trained on the large scale natural image database (ImageNet) and fine-tuned with T1CE and FLAIR images. The outputs of training set were utilized to train LR and SVM models. Besides, fused images combining FLAIR and T1CE were used to fine-tune pre-trained ImageNet models. Results: Performances were improved by fine-tuning the four architectures with fused images. Conclusion: Transfer learning with various CNNs (especially VGGNet) is powerful in predicting IDH1 genotype in grade Ⅱ/Ⅲ gliomas.